Its Unfitness to Deal with Increasing Returns...

The confusion lamented by Leijonhufvud derives from the habit of isolating Marshall’s analysis from its aim and context. Marshall bears some responsibility for this, because he stretched the mechanical analogy beyond the limits he had set and tried to shield it from its deficiencies, reserving their treatment to a further stage that he never achieved.

Increasing returns follow, among others, the pattern of the growth of knowledge as conceived in “Ye machine”: variations produce innovations that in the long run irre­versibly change the operating system. Book IV of the Principles and Industry and Trade show the model at work in shaping the organization of the industry, in a relentless suc­cession of innovation and standardization. This evolutionary model cannot be envisaged through the lens of static equilibrium analysis.

The issue of increasing returns troubled Marshall from the beginning. They were first introduced as part of the analytical machinery in curve D of the early essay “On value”, meant to explain long-term equilibrium. A letter to Neville Keynes shows that, when writing the Principles, Marshall was in doubt as to whether to resurrect the analysis of the early essay, initially set aside because of fear of “over-complexity” (Whitaker 1996, I: 278). In the first edition, increasing returns found their proper place in chapter V of book V, the core of value theory. Marshall’s fear of over-complexity must soon have had a comeback, as he felt unhappy with this collocation, and changed it in subsequent editions. Ultimately, in edition V, increasing returns were relegated to appendix H. The irreversibility problem was shut up in a sort of “Pandora’s box” (Bharadwaj 1972: 46), together with the limitations of the statical method that could have undermined the fun­damentals of equilibrium analysis.

Given the fact that irreversibility is an essential feature of increasing returns, they were a source of trouble, which, though it could not be ignored, was to be segregated, “softening the impact of the remark of the non-rigidity of the curves with the claim that its relevance was limited to the case of increasing returns only, with no substantial consequences for the other cases” (Dardi 2006: 224). Hence the decision to treat them in the appendix almost as if they were an anomaly. As Dardi observes, this decision leads readers astray, as it makes them believe that increasing returns are an exception to the rule that supply curves unproblematically provide a perfect representation of the eco­nomic forces at work. When Marshall faces the “exception”, in appendix H, the supply curves become two. The “true” supply curve plots the prices at which any quantity can be produced, assuming that quantity be the overall level of production. This curve ade­quately represents the case of increasing returns, with its intrinsic irreversibility. Beside it, Marshall drew the “particular expenses curve”, always positively inclined, which embodies the conditions under which different quantities of a commodity are produced under the assumption that total supply is fixed. Equilibrium along this curve coincides with the cost of the marginal producer. The “true” supply curve cannot be derived from the particular expenses curve. As Andrews (1951: 148) notes, “the reasoning is of general application”, and, here as elsewhere, “the blunt facts of increasing returns saved Marshall from the use of constructions which...

Increasing returns deprive the concept of margin itself of meaning:

[T]he term “margin of production” has no significance for long periods in relation to com­modities the cost of production of which diminishes with a gradual increase in the output... Therefore, when we are discussing the special conditions of value of those commodities which conform to that tendency, the term “margin” should be avoided (Marshall 1920: app. H).

To prevent the problem from disrupting the theory of the equilibrium of the individual firm, Marshall assumed that “a tendency to increasing returns does not exist generally for short periods” (ibid.). For the long period, the problem was dealt with by resort­ing to the representative firm, which has access to the internal and external economies dependent on the expansion of production (Marshall 1920: 317). The representative firm is meant to explain why internal economies do not lead to monopoly, as mathematical equilibrium theory inexorably infers from its premisses (Whitaker 1996, II: 227).

Though relegated to the appendix, increasing returns are not an oddity, but bring to the forefront the irreversibility problem, which is ubiquitous in Marshall’s economics. Hysteresis regards demand curves too, whenever new patterns of consumption, estab­lished by price variations, change the consumer’s preferences after prices have returned to their former level (Marshall 1920: 808; cf. also Whitaker 1975, II: 163). Such phenomena defy the tools of static equilibrium analysis. Given the mathematics he knew, Marshall could not do much better, but he never accepted the idea that statics has autonomous scientific status: “Statics is but a branch of Dynamics” (Marshall 1920: 366 n). The term dynamics itself, however, is deceptive, as it fails to seize the essential elements of the problem. While the analogy between economic and mechanical statics is fruitful, that between economic and mechanical dynamics is misleading: “the Mecca of the economist is economic biology rather than economic dynamics” (Marshall 1898: 43).

[I]n the earlier stages of economics, we think of demand and supply as crude forces pressing against one another, and tending towards a mechanical equilibrium; but in the later stages, the balance or equilibrium is conceived not as between crude mechanical forces, but as between the organic forces of life and decay. (Marshall 1898: 43)